{
  "best_metric": 5.5685906410217285,
  "best_model_checkpoint": "miner_id_24/checkpoint-50",
  "epoch": 0.13689253935660506,
  "eval_steps": 50,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0027378507871321013,
      "grad_norm": 2869.560302734375,
      "learning_rate": 1e-05,
      "loss": 18.9407,
      "step": 1
    },
    {
      "epoch": 0.0027378507871321013,
      "eval_loss": 5.184251308441162,
      "eval_runtime": 2.1423,
      "eval_samples_per_second": 287.075,
      "eval_steps_per_second": 71.886,
      "step": 1
    },
    {
      "epoch": 0.0054757015742642025,
      "grad_norm": 3299.45068359375,
      "learning_rate": 2e-05,
      "loss": 17.9501,
      "step": 2
    },
    {
      "epoch": 0.008213552361396304,
      "grad_norm": 2877.494140625,
      "learning_rate": 3e-05,
      "loss": 18.3079,
      "step": 3
    },
    {
      "epoch": 0.010951403148528405,
      "grad_norm": 3389.11767578125,
      "learning_rate": 4e-05,
      "loss": 19.2225,
      "step": 4
    },
    {
      "epoch": 0.013689253935660506,
      "grad_norm": 4282.4208984375,
      "learning_rate": 5e-05,
      "loss": 19.2442,
      "step": 5
    },
    {
      "epoch": 0.01642710472279261,
      "grad_norm": 3912.53466796875,
      "learning_rate": 6e-05,
      "loss": 18.8629,
      "step": 6
    },
    {
      "epoch": 0.019164955509924708,
      "grad_norm": 4039.3466796875,
      "learning_rate": 7e-05,
      "loss": 18.9577,
      "step": 7
    },
    {
      "epoch": 0.02190280629705681,
      "grad_norm": 5384.4765625,
      "learning_rate": 8e-05,
      "loss": 18.5148,
      "step": 8
    },
    {
      "epoch": 0.024640657084188913,
      "grad_norm": 3950.795654296875,
      "learning_rate": 9e-05,
      "loss": 18.7446,
      "step": 9
    },
    {
      "epoch": 0.02737850787132101,
      "grad_norm": 4083.84619140625,
      "learning_rate": 0.0001,
      "loss": 17.6605,
      "step": 10
    },
    {
      "epoch": 0.030116358658453114,
      "grad_norm": 3851.057861328125,
      "learning_rate": 9.999316524962345e-05,
      "loss": 19.1159,
      "step": 11
    },
    {
      "epoch": 0.03285420944558522,
      "grad_norm": 4973.96875,
      "learning_rate": 9.997266286704631e-05,
      "loss": 18.5552,
      "step": 12
    },
    {
      "epoch": 0.03559206023271732,
      "grad_norm": 3997.625732421875,
      "learning_rate": 9.993849845741524e-05,
      "loss": 18.5181,
      "step": 13
    },
    {
      "epoch": 0.038329911019849415,
      "grad_norm": 4821.0908203125,
      "learning_rate": 9.989068136093873e-05,
      "loss": 18.9077,
      "step": 14
    },
    {
      "epoch": 0.04106776180698152,
      "grad_norm": 7009.69091796875,
      "learning_rate": 9.98292246503335e-05,
      "loss": 18.1293,
      "step": 15
    },
    {
      "epoch": 0.04380561259411362,
      "grad_norm": 5173.1787109375,
      "learning_rate": 9.975414512725057e-05,
      "loss": 18.3298,
      "step": 16
    },
    {
      "epoch": 0.04654346338124572,
      "grad_norm": 5275.22705078125,
      "learning_rate": 9.966546331768191e-05,
      "loss": 17.488,
      "step": 17
    },
    {
      "epoch": 0.049281314168377825,
      "grad_norm": 6161.54345703125,
      "learning_rate": 9.956320346634876e-05,
      "loss": 18.125,
      "step": 18
    },
    {
      "epoch": 0.05201916495550993,
      "grad_norm": 5380.9541015625,
      "learning_rate": 9.944739353007344e-05,
      "loss": 19.6235,
      "step": 19
    },
    {
      "epoch": 0.05475701574264202,
      "grad_norm": 7076.9326171875,
      "learning_rate": 9.931806517013612e-05,
      "loss": 19.6551,
      "step": 20
    },
    {
      "epoch": 0.057494866529774126,
      "grad_norm": 9411.6328125,
      "learning_rate": 9.917525374361912e-05,
      "loss": 19.1774,
      "step": 21
    },
    {
      "epoch": 0.06023271731690623,
      "grad_norm": 5219.16845703125,
      "learning_rate": 9.901899829374047e-05,
      "loss": 17.0145,
      "step": 22
    },
    {
      "epoch": 0.06297056810403832,
      "grad_norm": 5389.611328125,
      "learning_rate": 9.884934153917997e-05,
      "loss": 18.715,
      "step": 23
    },
    {
      "epoch": 0.06570841889117043,
      "grad_norm": 4909.1533203125,
      "learning_rate": 9.86663298624003e-05,
      "loss": 18.0538,
      "step": 24
    },
    {
      "epoch": 0.06844626967830253,
      "grad_norm": 4762.9892578125,
      "learning_rate": 9.847001329696653e-05,
      "loss": 18.122,
      "step": 25
    },
    {
      "epoch": 0.07118412046543464,
      "grad_norm": 5046.6279296875,
      "learning_rate": 9.826044551386744e-05,
      "loss": 16.9833,
      "step": 26
    },
    {
      "epoch": 0.07392197125256673,
      "grad_norm": 5887.14208984375,
      "learning_rate": 9.803768380684242e-05,
      "loss": 18.4829,
      "step": 27
    },
    {
      "epoch": 0.07665982203969883,
      "grad_norm": 7543.8603515625,
      "learning_rate": 9.780178907671789e-05,
      "loss": 19.2867,
      "step": 28
    },
    {
      "epoch": 0.07939767282683094,
      "grad_norm": 4809.97705078125,
      "learning_rate": 9.755282581475769e-05,
      "loss": 18.0179,
      "step": 29
    },
    {
      "epoch": 0.08213552361396304,
      "grad_norm": 4606.07666015625,
      "learning_rate": 9.729086208503174e-05,
      "loss": 18.5569,
      "step": 30
    },
    {
      "epoch": 0.08487337440109514,
      "grad_norm": 4569.7158203125,
      "learning_rate": 9.701596950580806e-05,
      "loss": 18.9656,
      "step": 31
    },
    {
      "epoch": 0.08761122518822724,
      "grad_norm": 4704.12060546875,
      "learning_rate": 9.672822322997305e-05,
      "loss": 18.4432,
      "step": 32
    },
    {
      "epoch": 0.09034907597535935,
      "grad_norm": 5654.9912109375,
      "learning_rate": 9.642770192448536e-05,
      "loss": 18.5611,
      "step": 33
    },
    {
      "epoch": 0.09308692676249145,
      "grad_norm": 6667.42626953125,
      "learning_rate": 9.611448774886924e-05,
      "loss": 19.0092,
      "step": 34
    },
    {
      "epoch": 0.09582477754962354,
      "grad_norm": 9212.4580078125,
      "learning_rate": 9.578866633275288e-05,
      "loss": 19.2947,
      "step": 35
    },
    {
      "epoch": 0.09856262833675565,
      "grad_norm": 5465.796875,
      "learning_rate": 9.545032675245813e-05,
      "loss": 18.93,
      "step": 36
    },
    {
      "epoch": 0.10130047912388775,
      "grad_norm": 7496.95654296875,
      "learning_rate": 9.509956150664796e-05,
      "loss": 21.2021,
      "step": 37
    },
    {
      "epoch": 0.10403832991101986,
      "grad_norm": 8689.849609375,
      "learning_rate": 9.473646649103818e-05,
      "loss": 20.4252,
      "step": 38
    },
    {
      "epoch": 0.10677618069815195,
      "grad_norm": 7595.74365234375,
      "learning_rate": 9.43611409721806e-05,
      "loss": 21.0276,
      "step": 39
    },
    {
      "epoch": 0.10951403148528405,
      "grad_norm": 7750.20166015625,
      "learning_rate": 9.397368756032445e-05,
      "loss": 21.0263,
      "step": 40
    },
    {
      "epoch": 0.11225188227241616,
      "grad_norm": 9827.46875,
      "learning_rate": 9.357421218136386e-05,
      "loss": 21.9128,
      "step": 41
    },
    {
      "epoch": 0.11498973305954825,
      "grad_norm": 6157.37744140625,
      "learning_rate": 9.316282404787871e-05,
      "loss": 20.8637,
      "step": 42
    },
    {
      "epoch": 0.11772758384668036,
      "grad_norm": 7301.2265625,
      "learning_rate": 9.273963562927695e-05,
      "loss": 21.7574,
      "step": 43
    },
    {
      "epoch": 0.12046543463381246,
      "grad_norm": 7452.45654296875,
      "learning_rate": 9.230476262104677e-05,
      "loss": 21.4877,
      "step": 44
    },
    {
      "epoch": 0.12320328542094455,
      "grad_norm": 6658.0244140625,
      "learning_rate": 9.185832391312644e-05,
      "loss": 21.5988,
      "step": 45
    },
    {
      "epoch": 0.12594113620807665,
      "grad_norm": 9002.0439453125,
      "learning_rate": 9.140044155740101e-05,
      "loss": 21.8134,
      "step": 46
    },
    {
      "epoch": 0.12867898699520877,
      "grad_norm": 8484.044921875,
      "learning_rate": 9.093124073433463e-05,
      "loss": 22.6028,
      "step": 47
    },
    {
      "epoch": 0.13141683778234087,
      "grad_norm": 5393.41650390625,
      "learning_rate": 9.045084971874738e-05,
      "loss": 21.4441,
      "step": 48
    },
    {
      "epoch": 0.13415468856947296,
      "grad_norm": 7081.9052734375,
      "learning_rate": 8.995939984474624e-05,
      "loss": 21.2952,
      "step": 49
    },
    {
      "epoch": 0.13689253935660506,
      "grad_norm": 8458.6767578125,
      "learning_rate": 8.945702546981969e-05,
      "loss": 21.0256,
      "step": 50
    },
    {
      "epoch": 0.13689253935660506,
      "eval_loss": 5.5685906410217285,
      "eval_runtime": 2.1364,
      "eval_samples_per_second": 287.871,
      "eval_steps_per_second": 72.085,
      "step": 50
    }
  ],
  "logging_steps": 1,
  "max_steps": 200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 84793224069120.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}